At Twitter, to overcome performance issues that can arise in developing and maintaining SQL systems with increasing volumes of data, we designed a large-scale SQL federation system across on-premises and cloud Hadoop and Google Cloud Storage (GCS) clusters. leveraging Presto as the core of SQL engine clusters, pursuing high scalability and availability to fulfill the increasing need for data analytics on a petabyte (PB) scale of data. All the presto clusters have ran more than 10 million presto sql queries.